Nebius AI Studio
Nebius AI Studio offers serverless inference on a wide range of open models with a small free starter credit for new users, usable from playground or API after adding a bank card at signup.
- Free credit
- $1
- Credit window
- 30 days
- Card required
- Yes
Nebius AI Studio scores 5.7/10 for free value in LLM. The main constraint: New users receive a small free starter credit (reported $1) usable in the playground or via API, but a bank card is required at signup before onboarding completes; larger credits available via promos/programs. Finite.
Free value score
5.7 / 10Show the math
Key metrics
Models
40 free · 60+ totalLimits may be stale — last re-verified outside this section's freshness window.
| Model | Family | AA | Access |
|---|---|---|---|
| zai-org/GLM-5.2 | GLM | 51.1 | free |
| deepseek-ai/DeepSeek-V4-Pro | DeepSeek | 44.3 | free |
| moonshotai/Kimi-K2.6 | Kimi | 42.8 | free |
| zai-org/GLM-5.1 | GLM | 40.2 | free |
| zai-org/GLM-5 | GLM | 39.5 | free |
| moonshotai/Kimi-K2.5 | Kimi | 38.1 | free |
| nvidia/Nemotron-3-Ultra-550b-a55b | Nemotron | 37.8 | free |
| Qwen/Qwen3.5-397B-A17B | Qwen | 32 | free |
| deepseek-ai/DeepSeek-V3.2 | DeepSeek | 24.7 | free |
| openai/gpt-oss-120b | GPT | 23.8 | free |
| deepseek-ai/DeepSeek-V3.1 | DeepSeek | 21.1 | free |
| openai/gpt-oss-20b | GPT | 14.9 | free |
| nvidia/Llama-3_1-Nemotron-Ultra-253B-v1 | Llama | 9.1 | free |
| nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B | Nemotron | 7.4 | free |
| BAAI/bge-en-icl | BGE | — | free |
| BAAI/bge-multilingual-gemma2 | Gemma | — | free |
| deepseek-ai/DeepSeek-V3-0324 | DeepSeek | — | free |
| google/gemma-3-27b-it | Gemma | — | free |
| intfloat/e5-mistral-7b-instruct | Mistral | — | free |
| meta-llama/Llama-3.1-70B-Instruct | Llama | — | free |
| meta-llama/Llama-3.2-1B-Instruct | Llama | — | free |
| meta-llama/Llama-3.2-3B-Instruct | Llama | — | free |
| meta-llama/Llama-3.3-70B-Instruct | Llama | — | free |
| meta-llama/Llama-Guard-3-8B | Llama | — | free |
| meta-llama/Meta-Llama-3.1-8B-Instruct | Llama | — | free |
| MiniMaxAI/MiniMax-M2.5 | MiniMax | — | free |
| NousResearch/Hermes-4-405B | Llama | — | free |
| NousResearch/Hermes-4-70B | Llama | — | free |
| nvidia/Cosmos3-Super-Reasoner | Nemotron | — | free |
| nvidia/Nemotron-3-Nano-Omni | Nemotron | — | free |
| nvidia/nemotron-3-super-120b-a12b | Nemotron | — | free |
| openbmb/MiniCPM-V-4_5 | MiniCPM | — | free |
| PrimeIntellect/INTELLECT-3 | INTELLECT | — | free |
| Qwen/Qwen2.5-VL-72B-Instruct | Qwen | — | free |
| Qwen/Qwen3-235B-A22B-Instruct-2507 | Qwen | — | free |
| Qwen/Qwen3-30B-A3B-Instruct-2507 | Qwen | — | free |
| Qwen/Qwen3-32B | Qwen | — | free |
| Qwen/Qwen3-Coder-480B-A35B-Instruct | Qwen | — | free |
| Qwen/Qwen3-Embedding-8B | Qwen | — | free |
| Qwen/Qwen3-Next-80B-A3B-Thinking | Qwen | — | free |
Showing first 40 of a larger catalog.
What you get
Free starter credit (reported $1); bank card required at signup
Broad open-model catalog for text and image
OpenAI-compatible API and hosted playground
Cost-effective pay-as-you-go beyond the free credit
Gotchas
- $1 credit expires in 30 days
New accounts receive $1 in trial credit that is valid for only 30 days from sign-up and does not renew.
- $1 one-time credit pool
The free starter grant is a single finite $1 pool that never regenerates; sustained use requires adding paid funds.
- Bank card required at signup
Billing setup is mandatory and requires a bank card before onboarding can be completed, even to use the free trial credit.
- Larger credits need promo/program
Credit figures larger than $1 (e.g. $50 promo codes, $1000 contest prizes, research grants) come only from time-limited promotions or approval-gated programs, not the standard signup grant.
Sources
-
"Setting up a billing account requires a bank card. ... Billing setup is mandatory—you cannot complete onboarding without it. ... Upon first sign-up, you receive $1 in trial credit, valid for 30 days."
https://docs.tokenfactory.nebius.com/other-capabilities/billing-new -
"Baseline 1.00x | RPM limit 60 | TPM limit 400,000 ... The limit can grow to 20x your base allocation. Beyond that we require an Enterprise plan."
https://docs.tokenfactory.nebius.com/ai-models-inference/rate-limits -
"Large 10 point gain in Intelligence Index: DeepSeek V4 Pro (Max) scores 52 on the Artificial Analysis Intelligence Index ... making it the #2 open weights reasoning model behind Kimi K2.6 (54)."
https://artificialanalysis.ai/articles/deepseek-is-back-among-the-leading-open-weights-models-with-v4-pro-and-v4-flash -
"Nebius Token Factory is the next evolution of Nebius AI Studio, redesigned for enterprise readiness and full model-lifecycle management."
https://www.hpcwire.com/aiwire/2025/11/07/nebius-launches-token-factory-to-deliver-production-ai-inference-at-scale/
